Historical Bogo-Linguistics
or
How to change one language into something which looks and sounds like another
or
The Lazy Conlanging HOWTO

Geoff's homepage -> Artificial Languages -> Bogo-Linguistics

Last update: 14 September 2007


Introduction

"Historical bogo-linguistics"? What's that?

It's the name I've chosen for the method of conlanging whereby one language (the source) is given a phonology resembling that of another (the target). "Bogo" implies that, while fun, this is historically bogus. The term bogolang has been proposed (not by me!) as a classification for the resulting language, and I'm happy to support its offical adoption.

Why would anyone want to do something like that?

Explaining the appeal of conlanging is a bit like trying to explain jazz (niiiice); as [IIRC] Duke Ellington remarked, "If you gotta ask, you'll never know". However, assuming you understand why some people like making up languages, I'll explain the benefits of this particular approach, which I'll abbreviate as HBL to save my fingers.

The principal benefit of HBL is that it relieves the conlanger of one of the most tiresome aspects of creating a language, namely the need to generate a working vocabulary - assuming that the source language has enough of one already, thus Latin is probably a better choice than something like Gothic. (Although, with access to good Proto-Germanic word sources, I have managed to make a servicable bogolang out of Gothic). HBL is thus worth considering if you need to create a detailed conlang with a particular flavour quickly and aren't too bothered about it having its own unique lexicon.

Other benefits of HBL are: it creates words which look and sound vaguely familiar, but with meanings which won't be immediately obvious; it's a good way for fledgling conlangers to learn about language change and the conlanging process; and, in certain settings, it can inspire alternate historical timelines.

Ah, I see. Has anyone already done this?

The original and best example of HBL is Brithenig, which treats Latin as if it were Welsh; it's also part of a very convincing alternate timeline. In the same conworld are Kernu (Latin-as-Cornish), for which I am not aware of a Web presence, and my own Breathanach, which was inspired by Brithenig and develops Latin along the lines of Gaelic. There are also at least three Latin-to-Slavonic conlangs: Wenedyk, its close relative Slezan, and the currently rather less detailed Slvanjec.

What languages work best with HBL?

In theory, any two languages about which you know enough will do, as long as you can come up with sound-changes which work convincingly. Brithenig and Breathanach were helped considerably by the relatively straightforward conversion from the Latin phonology to the Celtic; by contrast, while a conversion from Arabic to Old English sounds very interesting, it would be quite a challenge.

The majority of bogolangs are romlangs, i.e. those which start from Latin; this is no doubt because so much is known about Latin and its subsequent evolution into the Romance languages. Langmaker's romlang page has a link to over fifty.

OK, I'm convinced. Can you show me what to do?

Sure. I'll describe how one might go about developing a Slavonic-flavoured romlang which might have been spoken somewhere in Eastern Europe in the early Middle Ages. I've chosen Slavonic not just because I know a lot about its phonology having already created the Slavonic-like Rachovian, but also because its phonological development is interesting and well documented. It is instructive to compare the results with Wenedyk, which takes a different approach to the conversion process.

There are four steps to the process of HBL:

  1. Choose a series of sound-changes which will convert the phonology of the source language into one resembling that of the target. This stage requires most of the hard work.
  2. Process the grammar and vocabulary of the source language through them. A program like my Sound Change Applier, or Mark Rosenfelder's "sounds" utility, is useful here.
  3. Tweak the sound-changes to get results more in line with some combination of your own prejudices preferences and actual historical developments.
  4. Repeat until you're satisfied with the results.

The conlang developed here is intended to be illustrative of the process, not a finished work of art. Consequently I've left in many rough edges and cut plenty of corners; anyone who wants to tidy it all up is welcome to do so.

Can I steal borrow what you've started here and finish it off?

Finally, a few words about copyright, or the lack of it. The conlang described here is placed in the public domain, which explains why I haven't bothered to think up a name for it; you may do what you like with it, as long as you do not restrict the ability of others to do what they like with it, and you credit me somewhere; something like "Inspired by Geoff's original efforts" will do fine :-)

Transcription conventions

Latin words are cited in UPPERCASE with vowel length indicated by an acute accent, thus VIDÉRE "to see". Words in the resulting conlang are cited in boldface in a phonemic transcription which resembles that of the Roman transliteration of Old Church Slavonic.


Choose your sound-changes

We will derive our sample conlang from Latin by using sound-changes like those which changed Indo-European (IE) into Common Slavonic (CS). This historical method obviously requires some knowledge of the history of your target language, but it will create something with more or less the correct internal phonological patterning. Alternatively, we could pretty much ignore history and use sound-changes which approximate to how a native speaker of the target language would unconsciously render the source language; this is the adaptive method which was chosen for Wenedyk, and it has the merit that the result resembles the source language more closely.

Ingredients

From the point of view of historical accuracy, it's better to start with Vulgar Latin (VL), which is what people actually spoke, rather than Classical Latin (CL); the main difference is that VL lost phonemic vowel length, whereas CL retained it. Unfortunately, while this was no problem for Brithenig, vowel length is important in the history of Slavonic; we will therefore need to start with CL here. I had the same problem with Breathanach, btw.

The phonemes of CL can be straightforwardly summarised as five short and five long vowels /i e a o u/ and /i: e: a: o: u:/, plus three diphthongs /ai au oi/ (of which /oi/ was very rare) and thirteen consonants /p b t d k g f s h m n l r/. To these may be added /j w/, the prevocalic allophones of /i u/.

The sound-system of CS was much more complicated. At one stage of interest, the vowels consisted of short /i e o u/ and long /i: e: a: u: y:/ (where /y:/ = [M:] or [1:]), plus short and long nasalised /a~ e~/. There were nearly thirty consonant phonemes, many of which were due to various processes of palatalisation:

TypeLabialDentalAlveolarPalatalPal-alvVelar
Voiceless stops /affricatespttst_jtSk
Voiced stops /affricatesbd d_j g
Voiceless fricatives  s s_jSx
Voiced fricatives vz z_jZ 
Nasals mn n_j  
Laterals  l l_j  
Trills  r r_j  
Semivowel    j  

Note that unlike CL, CS lacked /f/. This isn't actually a problem, since there's no compelling reason why our bogolang shouldn't have it.

Merger of back vowels

CS, along with Germanic, merged /o o:/ with /a a:/, reducing the five vowel qualities to four. Our conlang will do the same; the reflexes of BONUS and MALUS will thus have the same stem vowel (bàn, màl), and the reflexes of AMÓ and AMÁS will be indistinguishable as oma. This change will also merge the rare Latin /oi/ with the commoner /ai/.

Note that Wenedyk, following different principles, keeps /a/ and /o/ separate.

Simplification of geminates

Unlike IE, CL had many geminate (double) consonants. We will deviate from history and remove them from our conlang by reducing them to single consonants, lengthening the preceding vowel in compensation: MISSUM > mísą, ILLE > jíle. (This dodge could perhaps be used to reintroduce vowel length into VL, if someone feels like it.)

A minor relevant detail is that intervocalic /s/ was voiced to [z] in Romance; this may be useful, but is ignored here.

Ruki-backing

As in CS, /s/ not followed by a stop will become /x/ after /r u k i/. This /x/ will be transcribed h to prevent confusion with Latin /ks/ which of course is written X. Thus FÚSUS > fýh and URSA > urha, later rho. (Note that the /r/ is syllabic; this word would be pronounced /r.xo/.)

Latin /h/ normally disappeared in the Romance languages. In our conlang, however, we will cheat slightly by merging it with /x/: HORTUS > hràt, pronounced /xra:t/.

The satem change

In CS, as well as Sanskrit, PIE /k g/ became /s z/, while /kw gw/ became /k g/. It is not clear whether to include this change in our conlang; /gw/ is rare in Latin, occurring only after /n/, and /g/ would thus be limited to very few words. On the other hand, keeping /kw/ would mean that the horribly Esperantine cluster kv would be uncomfortably frequent. We will compromise by changing /kw/ to /k/, as in French, but otherwise ignoring the satem change.

The first palatalisation

This is one of the most characteristic sound-changes in CS, although something similar happened in many other languages; it caused /k g x/ to become /tS dZ S/ before front vowels, with /dZ/ later simplifying to /Z/. This is easily interesting enough to use in our conlang, and we will transcribe the resulting sounds in the standard Slavonic manner as č ž š: GENESTA > ženesto, CENTUM > čętą.

Simplification of certain clusters

We will resolve /gn/ to /nj/, a common change in many Romance languages which also occurred in CS. Thus IGNIS > jìń.

/kt/ will become /st_j/: OCTÓ > oste /ost_je/.

Post-jod fronting

A phonetic process in CS with the unwieldy name of "the principle of intrasyllabic harmony" fronted the back vowels /a a: u u:/ to /e e: i i:/ after /j/ (strictly speaking, after any soft consonant). Thus ALIUM > olę and FILIÁ > fíle, by contrast with OVUM > ovą and PORTÁ > prota.

In our conlang, this will theoretically create two new diphthongs /ei/ and /eu/ from /jai/ and /jau/, although I doubt that /jau/ actually existed in Latin.

In CS, /e:/ became /a:/ after /j/. We will ignore this detail here.

Jodation

Combinations of consonant + /j/ were frequent in both CL and IE. As with the first palatalisation, the results will be included in our conlang:

/j/ as an independent phoneme thus only remains between vowels. The palatal sounds will also develop as allophones of the dentals before /e i/, as will /s_j z_j/ (otherwise transcribed ś ź) for /s z/; in other words, ti te represent /t_ji t_je/, not /ti te/. /s/ will become /S/ before /C/ and /s_j/ before palatals.

Reduction of diphthongs

At this stage CS had four diphthongs /ei ai eu au/, which simplified to /i: e: ju: u:/. These long vowels were pronounced with a falling tone distinct from the rising tone on the original long vowels. The same developments will occur in our conlang; the tonal distinction will be represented in transcription here in the traditional way, with acute accents for the rising tones and circumflex for the falling. Thus LAETA > lêto and TAURUS > tûr.

Unrounding of long high back vowels

The lack of distinctive rounding in back vowels meant that long /u:/ would have been weakly rounded, if at all. The collapse of /au/ to a fully rounded /u:/ caused this vowel to unround altogether to [M], here transcribed y: DÚRUS > dýr.

Syllable opening

One notable characteristic of CS was its preference for open syllables, which resulted from the elimination of all syllable-final consonants.

Nasal consonants left their nasality in the preceding vowel: front vowels gave nasal /e~/, transcribed ę, and back vowels gave nasal /a~/, transcribed ą. For example: CANTÁRE > kątare, PENSUM > pęsą.

Syllable-final liquids in CS developed in a variety of ways. This is altogether too complicated for our conlang, and we will merely follow the general principle whereby they change places with the preceding vowel: PARTEM > protę, ALNÍ > loni. [A more Russian-like sound-change, exhibiting a phenomenon known as polnognasie (sp?), would create porotę and oloni.] Liquids will simply be lost at the ends of words.

The only syllable-final fricative in Latin was /s/. In our conlang it will be transferred to the beginning of the following syllable, as with GENESTA > ženesto.

This leaves just the stops, which will be eliminated altogether: SEPTEM > setę, REXÍ > reši.

Prothetic consonants

The preference for open syllables led to a tendency to prefix a glide consonants to a word-initial vowel. In CS this caused a /j/ to appear before initial front vowels and /a:/, and /w/ before /u y/. So with our conlang: AUDÍRE > ûdíre > wûdíre > vûdíre; ESTIS > eśtI > jèśť.

Related to this is the hardening of prevocalic /w/ to /v/, as with vûdíre above.

The second and third palatalisations

In CS, /k g/ became /ts dz/ before the /e:/ from /ai/ (the second palatalisation), and after /i/ (the third). /x/ variously became /S/ or /s_j/; we will choose /S/, which is otherwise not very frequent. As with /dZ/, /dz/ often simplified to /z/.

In our conlang, we will extend the occurence of the third palatalisation to all instances after a front vowel, and transcribe /ts/ as c: CAEDÉRE > cêděre, TEGÓ > teza, TRAHEÓ > troše.

Change in quality of short vowels

This is the most complicated stage in the development of the CS vowels. At this stage the vowel system of our conlang contains long /i: e: a: u: y:/, short /i e a u/, and nasal /ę ą/; the short vowels, as in CS, will be assumed to have been pronounced more laxly and centrally than their long correspondents. Thus short /e a/ would become [e o], with long /e:/ being opener, more like [æ:]. This long /e:/ is transcribed ě in non-initial syllables.

Short /i u/ will, as in CS, reduce to schwa-like sounds known as the jers, here transcribed I U. These will eventually disappear altogether, except before another jer, lengthening the preceding vowel and giving it a new rising tone called the neoacute, here represented with the grave accent (as in Rachovian, although it really should be a tilde): PISCIS > pišči > pìšč; ALIUS > aľis > oľI > òľ.

The result

At this stage, which is a good point at which to finish with the sound-changes, we have managed to recreate the CS consonant-system, with the addition of /f/. The vowel system of our conlang contains eight oral phonemes /i I e a o U u y/, of which /i e a y/ could be long and acute, /i e u/ long and circumflex, /I e a U/ long and neoacute, and /I e o U/ short; and two nasal vowels /a~ e~/.

To simplify the vowel system, and reduce the number of accents, the tonal distinctions will be eliminated everywhere except in initial syllables, and long vowels will be shortened word-finally.

Stress

The location of the stressed syllable in CS was bound up with tone, and was a very tricky topic further complicated by the different preferences in modern Slavonic: initial in Czech, penultimate in Polish, and all over the place in Russian and Serbo-Croat. We will ignore the stress here, because it's too complicated and not actually important unless our conlang expects to be spoken.

Develop the source language

Now that we've worked out a set of sound changes, we will subject Latin grammar to them and see what comes out. The file I used with my Sound Change Applier is available here.

Nouns

As in Vulgar Latin, we'll simplify things by assuming that fourth and fifth declension nouns joined the second and first declensions respectively. We'll start with the first declension noun PORTA "gate", whose declension turns out to look like this:

CaseLatinSingularLatin Plural
Nomporta proto portae prote
Accportamprotąportásprota
Genportaeprote portárumprotarą
Datportaeprote portís proti
Ablportá prota portís proti

A velar consonant at the end of the stem will undergo the first two palatalisations: TOGA TOGAE TOGÍS > togo tože toži.

So far, so ordinary. Now, look what happens with a noun in which the final /a/ is preceded by /j/, such as FAGEA "beech":

CaseLatinSingular Latin Plural
Nomfágea fáže fágeae fáži
Accfágeamfážęfágeás fáže
Genfágeaefáži fágeárumfážerą
Datfágeaefáži fágeís fáži
Ablfágeá fáže fágeís fáži

Except in the dative and ablative plural, the endings are different, thanks to the fronting of the thematic vowel /a/ after the preceding /j/. This alternation of endings rather nicely reproduces a similar phenomenon in Slavonic, which is after all the whole point of using Slavonic sound-changes.

The second declension nouns PORCUS "pig" and GLADIUS "sword" would decline as follows:

Case Latin Singular Latin Plural
Nom porcuspràk porcí proči
Acc porcumproką porcós proka
Gen porcí proči porcórumprokarą
Dat/Ablporcó proča porcís proči

Case Latin Singular Latin Plural
Nom gladiusglàď gladií glodi
Acc gladiumglodęgladiós glode
Gen gladií glodi gladiórumgloderą
Dat/Ablgladió glode gladiís glodi

Here, as well as the alternations of vowels and consonants, there's an internal alternation in which the last vowel of the stem is long in the nominative singular, due to the loss of a final jer, and short elsewhere.

Third declension nouns constitute a much more heterogenous collection in Latin, often with one syllable fewer in the nominative singular than in the other cases. This syllable will usually be restored analogically in our conlang. Here are the declensions of HOMÓ "man" and RÉX "king".

CaseLatin SingularLatin Plural
Nom homó hom[n]ahominés hàmne
Acc hominemhàmnęhominés hàmne
Gen hominishomìnhominum hàmną
Dat hominí hàmnihominibushàmnib
Abl homine hàmnehominibushàmnib

CaseLatinSingularLatin Plural
Nom réx ré[ž] régés réže
Acc régemréžęrégés réže
Gen régisréž régum rézą
Dat régí réži régibusréžib
Abl rége réže régibusréžib

No modern descendant of Latin preserves anything like as much as this nominal morphology, and in our conlang we'll accordingly have to get rid of some of the cases. An obvious first candidate for elimination is the genitive, which in virtually all Romance languages was expressed with de and the ablative; we can thus replace proči with de proča. Similarly, the dative and ablative are frequently the same, and we can use the identity of the nominative and accusative plural in the third declension to get rid of the accusative (and thus the need to type all those accents!); finally, there's no need to keep the final b in the third declension dative and ablative plurals. [Purists will, correctly, object that the accusative, rather than the nominative, was retained in Romance.]

The result is a two-case system reminiscent of Romanian or Old French, in which the oblique case represents the older dative and ablative, and could be used to express different meanings with some prepositions: ę pràk "into the pig", ę proča "in the pig". There are five separate patterns of inflection (i.e. declensions):

DeclensionNom singObl singNom plurObl plur
1 < 1 hardprotoprotaproteproti
2 < 1 softfážefážefážifáži
3 < 2 hardpràkpročapročiproči
4 < 2 softglàďgloďegloďigloďi
5 < 3 homahàmnehàmnehàmni

First and second declension nouns are largely feminine, third and fourth declension nouns masculine, and those of the fifth declension can be either.

I've ignored neuter nouns, which generally became masculine in Romance, up to now. If we wish to retain them, they would decline like third and fourth declension nouns, but with nominative singular in -ą -ę and nominative plural in -a -e. bàklą "stick" and olę "garlic" would thus decline like this:

DeclensionNom singObl singNom plurObl plur
Neuter hardbàkląbàklabàklabàkli
Neuter softolęoleoleoli

Adjectives

Adjectival morphology in Latin is much the same as nominal morphology; all we need to worry about are the inflections for gender in addition to case and number. Here is how BONUS "good", TERTIUS "third", and GRAVIS "heavy" decline.

GenderNom sing Obl singNom plurObl plur
Masc bàn bona boni boni
Fem bono bona bone boni
Neut boną bona bona boni
Masc trèť trete treti treti
Fem trète trete treti treti
Neut trètętrete trete treti
Masc/Femgràv grove grovle grovi
Neut grovęgrove grove grovi

In our conlang, as in VL, we will express comparisons with the reflexes of Latin MAGIS and MINUS, viz. màž and mìn. Thus màž gràv "heavier".

Demonstratives

Deriving the words for "this", "that" and "yonder" from ISTE, IPSE and ILLE, as in Italian, we get jste jse jile. We can keep those initial /j/'s before consonants, change them to /ji/, or drop them altogether; the last option looks the nicest, and reflects what actually happens in spoken Czech.

Adverbs

We will use the characteristic Romance compounds with MENTE and the feminine singular to express adverbs: grovemęte "heavily". grovmęte, with loss of the final vowel of the adjective, looks interesting, too.

Pronouns and articles

Personal pronouns tend to hold onto case-inflections which nouns and adjectives lose. With this in mind, our conlang will retain four cases for pronouns: nominative, accusative, genitive, and oblique.

Theoretically, the first person singular pronouns would derive from Latin EGO, ME, MEUS and MIHI. This is no problem for the accusative and oblique, which come out as me and mìš. However, jezo for "I" is unwieldy; let's give it a spoken reduced form zo. The genitive MEUS is syntactically an adjective; we'll pretend that it was pronounced /mejus/, which will become mèj and decline like trèť.

The second person singular is simpler. Nominative TÚ and accusative TE simply become ty te, and oblique TIBI becomes tìb, which doesn't look right and will be changed to tìš by analogy with the first person. Genitive TUUS is slightly more complicated; by converting it to /tuwus/ we get tùv tvo, which declines like bàn.

The first and second person plural pronouns, which are very similar, can be treated together. Nominative and accusative are alike, viz. NÓS VÓS, and come out as na va; the oblique NÓBIS VÓBIS develop into náb váb, which will be changed analogically as with the second person singular. The genitive NOSTER NOSTRÍ NOSTRA NOSTRAE becomes nostre nostri nostro nostre, with nostre by analogy instead of the expected noste; VOSTER is similar, although some might prefer to use the forms from VESTER instead.

The third person pronouns are more tricky. We'll follow the example of most Romance languages and derive them from the demonstrative ILLE; this will produce the nominative singulars jíle jílo jíl and plurals jíli jíle jíla. The accusatives will be le lo lU and li le la, derived from shortened forms of these. The genitives will follow French and Italian; in the singular these will derive from SUUS and parallel the second person singuar: sùv svo. The plural will derive from (IL)LÓRUM (IL)LÁRUM, both of which give làrą, which will be the same regardless of gender and number. This leaves the oblique, which will be analagous to the others.

In summary, our personal pronoun system looks like this:

Person Nom AccGenObl
1 sing zo memèj mejemìš
2 sing ty tetùv tvo tìš
3 sing mascjílelesùv svo sìš
3 sing fem jílolosùv svo sìš
3 sing neutjíl lUsùv svo sìš
1 plur na nanostre nostronáš
2 plur va vavostre vostrováš
3 plur mascjílililàrąliš
3 plur fem jílelelàrąleš
3 plur neutjílalalàrąleš

[I don't much like the way that in the third person the feminine plural is the same as the masculine singular. There are two ways around this, if it's a problem: use the masculine plural for both, or tweak our sound changes so that /e:/ becomes /ja:/ word-finally. This second solution would have the interesting effect of using the same four vowels /a e i o/ to indicate gender and number as Italian, but with different assignations. It would also, unfortunately, mess up other parts of the grammar.]

Our interrogative and relative pronouns will descend from QUI[S] QUEM QUOD; thus či "who", čę "whom", ko "what". The genitive "whose" is kýj from CUIUS. The oblique can follow the pattern of the personal pronouns: čiš koš.

Indefinite pronouns in Latin are compounds with QUI[S], and so with our conlang: ALIQUIS "someone" > olič, ALIQUOD "something" > àlco.

Prepositions

There's no point in listing a full set of prepositions here, so we'll content ourselves with the following selection.

Verbs

Verbal morphology, the most intricate part of Latin grammar, is always the most fun part of a romlang. We will give our conlang a verbal system similar to those of the contemporary Romance languages.

Present indicative

Here are the expected outcomes of one verb from each of the four Latin conjugations:

PersonAMÁREHABÉREMITTEREDORMÍRE
1 singoma hoble míta dromle
2 singoma hobe míť dromi
3 singomo hobe míť dràm
1 pluromam hobem mítim dromim
2 pluromaťhobeťmítiťdromiť
3 pluromą hobę mítą dromlę

There are some interesting alternations here, particularly in the fourth conjugation; but straightaway we can see potential problems with the ambiguities in the singulars of the first three conjugations. There are three ways out of this:

  1. Use subject pronouns to disambiguate; French does this, as do Brithenig and Breathanach, for the same reasons. Thus zo oma "I love", ty oma "you love". This is probably overkill for our conlang, though.
  2. Redo the sound-changes. For example, final /s/ could raise and lengthen a preceding vowel, giving omy hobi miti dromi. This would solve this problem, but would also mean redoing much of the nominal morphology; in particular, all second declension nouns would end in -y or -i in the nominative singular, which creates problems of its own.
  3. Change some endings by analogy. This is tricky, since it's not obvious which way the analogy should operate; probably the nicest solution is to borrow the fourth conjugation 2 singular -i for the other conjugations, thus omi "you love". This is the best solution; it looks a bit Italian, but there's nothing wrong with that.

Imperfect indicative

This is straightforward: for AMÁRE the expected forms are omabą omaba omabo omabam omabať omabą, and the first person singulars of the other three are hobebą mitebą dromibą. [Wenedyk follows the example of Vulgar Latin, and uses /v/ here where we have /b/.]

Note the homophony between the first singular and third plural. If this is a problem, it can be eliminated by changing the first and second singulars to omaba omabi by analogy with the present.

Perfect indicative

This tense requires some thought, not just because the endings are different, but also because we have two sets to choose from in the first and fourth conjugations. For example, in the first conjugation, do we use the full CL form AMÁVÍ, or the reduced VL form AMÁÍ? Let's try it both ways:

CLomaviomavstiomav omavímomavisťomaverą
VLomeoměstiome oměmoměsťoměrą

The results with VL look more interesting and less obviously like CL, so we'll use those. Thus we get the following for our four sample verbs:

PersonAMÁISSEHABUISSEMÍSISSEDORMÍISSE
1 singome hobvi míši dromi
2 singoměstihàbvstimísistidromisti
3 singome hàbv míš dromi
1 pluroměm hobvìm míšìm dromim
2 pluroměsťhobvisť mísisťdromisť
3 pluroměrąhobvěrą míšerądromirą

This is almost fine; to tidy it up we will invoke analogy again, change ome to omi in the 1 singular, and make the 3 singular ending -e in all conjugations. There's also the matter of the rather bizarre hàbvsti; it's probable that this would retain its jer and remain habvisti. A better and neater solution is to get rid of the /v/ in the second conjugation altogether, thus hobi habisti habe, etc.

Note that t in the second person singular represents /t_j/.

Future and conditional indicative

The future tense in most Romance languages derives from a combination of the infinitive and the present tense of HABÉRE, thus French aimerai, Spanish amaré. The conditional similarly used the imperfect of HABÉRE: aimerais, amaria. For our conlang this is not practical since the sound-changes don't give us a sufficiently reduced auxiliary, and we'd have to contrive something like omarble for the future and omarbeba for the conditional. Instead, and perhaps more reasonably, we will follow the example of Romanian and use the descendant of Vulgar Latin VOLÉRE (conjugated like HABÉRE) as an auxiliary: vole omare "I will love", voleba omare "I would love".

Subjunctives

The two tenses of the subjunctive mood which survived into Romance were the present and pluperfect; this latter became the imperfect subjunctive with the characteristic -ss- in French and Italian.

The present subjunctive was formed by changing the thematic vowel. The first attempt for our sample verbs comes out like this:

PersonAMÁREHABÉREMITTEREDORMÍRE
1 singomąhoblę mítądromlę
2 singome hoble míta dromle
3 singome hoble míto dromle
1 pluromem hoblem mítam dromlem
2 pluromeťhobleťmítaťdromleť
3 pluromę hoblę mítą dromlę

This is pretty much OK as is; there are some homphonies, but nothing too troubling.

The imperfect subjunctive is easy, subject to adjustments like those for the perfect indicative:

PersonAMÁISSEHABUISSEMÍSISSEDORMÍISSE
1 singomasęhobisęmísisędromisę
2 singomasehobisemísisedromise
3 singomasehobisemísisedromise
1 pluromasemhobisemmísisemdromisem
2 pluromaseťhobiseťmísiseťdromiseť
3 pluromasęhobisęmísisędromisę

Infinitives and participles

All of these non-finite parts of the verb are straightforward. The present infinitives resemble their Latin ancestors: omare hobere mítere dromíre. The perfect infinitive did not survive into Romance; however, if we wanted one, it would replace the -re of the present infinitive with -se. We could, if we wanted, lose the final -e from both infinitives, which would produce omaŕ omaś and so on.

The present participles are omąte hobęte mítęte dromęte, and the past participles are omat hobet mís dromit. mís might as well become míst by analogy, and similarly for other past participles in -s such as klûht "closed" (with ruki-backing; from CLAUSUS). Note also participles in -sť, such as tésť "covered" from TÉCTUM.

With the past participle and the appropriate forms of "to have" we can form the familiar Romance present perfect and pluperfect tenses: hoble omat "I have loved", hobebą omat or hobi omat "I had loved". The future perfect is thus vole hobere omat "I will have loved".

Irregular verbs

The principal irregular verb is "to be", which when derived directly from the Latin comes out as są e e sùm jèśt są. The homophony between the 1 singular and the 3 plural doesn't trouble speakers of Italian, and it shouldn't really worry us either. However, no Romance language (except perhaps French) has homophony between the 2 and 3 singular; to resolve this, we'll follow the example of Italian and prepend /s/ in both second person forms, giving se sèśt. Note that the initial s is pronounced /s_j/, since it appears before the front vowel /e/.

The imperfect of "to be" derives from ERAM and so on: jerą/jera jera/jeri jero jeram jerať jerą. The perfect comes from FUÍ, and we will simplify the theoretical fvi to the more genial fi.

"To be" can be combined with the past participle to make the passive: są omat "I am loved", fi omat "I was loved". Its own participles wil be derived from STÁRE.

Most classically "irregular" Romance verbs are either the result of suppletion (as with "to be" in our conlang) or irregularities caused by sound change in common verbs (e.g. French j'ai, tu as, il a). The second type of irregularity is visible in our conlang in the present singulars of "to please" (ploko plači plač) and "to say" (díco díči dič); these alternations are of course due to the first and third palatalisations.

Summary

Here are the revised conjugations of all four verbs and "to be".

Present indicative

PersonAMÁREHABÉREMITTEREDORMÍREESSE
1 singoma hoble míta dromle
2 singomi hobi míti dromi se
3 singomo hobe míť dràm e
1 pluromam hoběm mítìm dromim sùm
2 pluromaťhoběťmítìť dromiťsèśt
3 pluromą hobę mítą dromlę

Imperfect indicative

PersonAMÁREHABÉREMITTEREDORMÍREESSE
1 singomaba hobeba míteba dromiba jera
2 singomabi hobebi mítebi dromibi jeri
3 singomabo hobebo mítebo dromibo jero
1 pluromabamhobebammítebamdromibamjeram
2 pluromabaťhobebaťmítebaťdromibaťjerať
3 pluromabąhobebąmítebądromibąjerą

Perfect indicative

PersonAMÁISSEHABUISSEMÍSISSEDORMÍISSEESSE
1 singomi hobi mísi dromi fi
2 singoměstihàbistimísistidromístifisti
3 singome hàbe míšedromefe
1 pluroměmhobimmíšimdromimfìm
2 pluroměsťhobisť mísisťdromisťfisť
3 pluroměrąhoběrą míšerądromirąfirą

Present subjunctive

PersonAMÁREHABÉREMITTEREDORMÍREESSE
1 singomąhoblęmítą dromlę
2 singome hoble míta dromlesi
3 singome hoble míto dromlesi
1 pluromem hoblem mítam dromlemsìm
2 pluromeťhobleťmítať dromleťsìť
3 pluromę hoblęmítą dromlę

Imperfect subjunctive

PersonAMÁISSEHABUISSEMÍSISSEDORMÍISSEFUISSE
1 singomasęhobisęmísisę dromisęfísę
2 singomasehobisemísisedromisefíse
3 singomasehobisemísisedromisefíse
1 pluromasemhobisemmísisemdromisemfísem
2 pluromaseťhobiseťmísiseť dromiseťfíseť
3 pluromasęhobisęmísisę dromisęfisę

Others

AMÁREHABÉREMITTEREDORMÍREESSE
Infinitiveomarehoberemíteredromírejésere
Present partomątehobętemítęte dromętestąte
Past partomathobetmístdromitstát


What next?

Here are a few suggestions for further sound-changes, in case anyone's interested.

Finally, the tones could be eliminated, as in most contemporary Slavonic languages except Slovene and Serbo-Croat. One possibility is for neoacute vowels to remain long, with either the acute or circumflex (or even both!) vowels shortening.


Samples

Our bastard offspring of Latin and Slavonic may currently be only a sketch; but the results, as is often the case with HBL, are quite piquant; the preference for open syllables means that it sounds a bit like Italian. At this point we'd normally now do a Babel text; for a mere demonstration like this, however, a few samples will suffice. It's interesting to compare these with Wenedyk.

The numbers up to ten look like this:

  1. ÚNUS ÚNÁ ÚNUM výn výno výną. Here, as elsewhere, there's the curious (for a Romance language) use of -o for the feminine, not the masculine.
  2. DUÓ DUÁ: dva.
  3. TRÉS TRIA: tre
  4. QUATTUOR: kátvo.
  5. QUÍNQUE: čęče.
  6. SEX: se, maybe seś.
  7. SEPTEM: setę.
  8. OCTÓ: oste, perhaps ostę by analogy.
  9. NÓVEM: návę.
  10. DECEM: dečę.

And the months come out as:

Finally, a full sentence; the rather approriate "you can lead a horse to water, but if you can make him float on his back, you've got something". My translation of the Latin is equum aquam dúcere potes, sed si facere ut super dorsum innat, causam bonam habes; in our conlang it looks like this, using contemporary Romance syntax:

poti dyčere jèc od oko, se si poti le fočere spe sva droha, hobi ýno kûho bono.

Groovy, isn't it? If you like, you can respell it as if it were Polish:

poci dyczerze jec od oko, sie si poci le foczerze spe swa drocha, chobi yno kucho bono.